The study investigated the effect of object preference on the learning of signs by 12 severely and profoundly retarded children (4-7 years old). Signs were chosen that were two-handed, symmetrical, and abstract. Training of each sign consisted of 10 individual training sessions using operant techniques. Results revealed no differences in the learning of signs for preferred versus nonpreferred objects. Three possible reasons for the findings (small sample size, none of the objects were most highly valued, and the use of operant techniques) are offered. (Author/CL)